Patterns in Global Terrorism
Overview of Trends
We start by plotting the overall number of attacks, along with the number of death and injury incurred.
Below we map every single terrorist attack recorded in the GTD for the years 2013, 2014, and 2015.
Here we break down the total number of attacks per year by the type of the attack. We note that bombings have increased quite a lot relative to other types of attacks.
We do the same for the types of targets that were victim to these attacks. As the plot shows, private citizens & property has experienced the largest relative increase.
Interstingly, suicide attacks experienced “peak popularity” during the height of the Iraq War in 2004 - 2006.
Patterns Across Terrorist Organizations
We observed high number of incidents after 2010. Here we random picked four years to see whether there is any pattern within a year. According to the plot, there is no obvious peak for specific month for all years. This is reasonable because if the terrorism patterns are consistent over years, the police will stay alert during the days that attacks have a highly chance to happen.
Finding that top 10 terrorist organization already account for 20% of the total attacks.
## Loading tidyverse: tibble
## Loading tidyverse: tidyr
## Loading tidyverse: readr
## Loading tidyverse: purrr## Conflicts with tidy packages ----------------------------------------------## annotate(): ggplot2, NLP
## filter(): dplyr, plotly, stats
## lag(): dplyr, stats
So we want to focus more on the top terrorist organizations to draw more conclusions. In order to determine who are the top gangs, we made an assumption to define the importance of a particular organization:
- It is very active (has large number of attacks)
- Its attacks have high fatality rate
Here is the listing of these top offenders:
We can see that these three winner gangs perform their attack in different styles. Taliban has the most attacks throughout the years; Dissident Military Members of Northern Tribal Group has just one attack but killed 150 people with none survived; Aum Shinri Kyo has several attacks, wounding many people but not causing heavy fatal consequences.
We filter out 9 gangs which qualifies for “big gangs”. We find that although they only account for 1% of total terrorist attacks, they are responsible for 3% of total death and 2% of total wounded.
We can see that these top gangs each has a territory of its own.
With this plot we see a clear expansion of ISIS, from limited to Iraq to spreading to Afraica and eventually France.
Text Analysis
These are the most common terms used to describe the motive of a terrorist attack in the GTD.
In the 1970s, the most common terms looked something like this.
This is what the most common terms looked like in 2015.
We also take a look at some interesting trends in the frequency of words over time.
Finally, we can also look at the most popular terms used to describe the motives of the largest terrorist organizations.